Journal of

Marine Science
and Engineering

Article
Hydrodynamic Zone of Influence Due to a Floating
Structure in a Fjordal Estuary—Hood Canal Bridge
Impact Assessment
Tarang Khangaonkar * , Adi Nugraha and Taiping Wang
Pacific Northwest National Laboratory, U.S. Department of Energy, Seattle, WA 98109, USA;
adi.nugraha@pnnl.gov (A.N.); taiping.wang@pnnl.gov (T.W.)
* Correspondence: tarang.khangaonkar@pnnl.gov; Tel.: +1-206-528-3053




Received: 17 September 2018; Accepted: 10 October 2018; Published: 15 October 2018


Abstract: Floating structures such as barges and ships affect near-field hydrodynamics and create
a zone of influence (ZOI). Extent of the ZOI is of particular interest due to potential obstruction to
and impact on out-migrating juvenile fish. Here, we present an assessment of ZOI from Hood Canal
(Floating) Bridge, located within the 110-km-long fjord-like Hood Canal sub-basin in the Salish Sea,
Washington. A field data collection program allowed near-field validation of a three-dimensional
hydrodynamic model of Hood Canal with the floating bridge section embedded. The results confirm
that Hood Canal Bridge, with a draft of 4.6 m covering ~85% of the width of Hood Canal, obstructs the
brackish outflow surface layer. This induces increased local mixing near the bridge, causes pooling of
water (up-current) during ebb and flood, and results in shadow/sheltering of water (down-current).
The change in ambient currents, salinity, and temperature is highest at the bridge location and reduces
to background levels with distance from the bridge. The ZOI extends ~20 m below the surface and
varies from 2–3 km for currents, from 2–4 km for salinity, and from 2–5 km for temperature before the
deviations with the bridge drop to <10% relative to simulated background conditions without the
bridge present.

Keywords: Hood Canal; floating bridge; Salish Sea; hydrodynamics; Finite-Volume Community
Ocean Model (FVCOM); circulation; anthropogenic impact; zone of influence; Salish Sea model

1. Introduction
Hood Canal is a fjordal sub-basin within the Salish Sea region of Pacific Northwest. The Salish
Sea is a collective name given to waterbodies that include the Strait of Juan de Fuca, Strait of Georgia,
Puget Sound, and all their connecting channels and adjoining waters (see Figure 1). In the spring and
summer, many Salish Sea sub-basins regularly experience algae blooms, and some of the sub-basins
such as Hood Canal, East Sound, and regions of South Sound show signs of hypoxia [1]. Development
of a comprehensive water-quality model of the Salish Sea was initiated in response to the above
concerns about management of nutrient pollution and the assimilative capacity of the Salish Sea [2–4].
Of particular interest was the ability to simulate low dissolved oxygen (DO) events in the Salish
Sea, including those responsible for fish kills and other chronic impacts in the Hood Canal region.
Numerous studies were conducted in the past to determine what contributes to low DO events in
Hood Canal, such as natural meteorological and oceanographic conditions, as well as anthropogenic
causes, such as excessive nutrient loading [5,6].
The Hood Canal sub-basin includes a unique anthropogenic modification in that it hosts the Hood
Canal Bridge (HCB), which is one of 11 floating bridges in use in the world. The geographic location
of this 1992-m-long bridge, about 10 km from the mouth of Hood Canal, is indicated in Figure 1.

J. Mar. Sci. Eng. 2018, 6, 119; doi:10.3390/jmse6040119 www.mdpi.com/journal/jmse

J. Mar. Sci. Eng. 2018, 6, 119 2 of 22
J. Mar. Sci. Eng. 2018, 6, x 2 of 22

1. Hood
Hood Canal
Canal is ais110-km-long
a 110-km-long narrow
narrow fjord-like
fjord-like sub-basin
sub-basin of of Puget
Puget Sound
Sound with
with anan averagewidth
average widthof
of 2.4
2.4 km,km, a mean
a mean depth
depth of of 51.1
51.1 m,m, anda amaximum
and maximumdepthdepthofof~187
~187 m,m, and it has
has aa sill
sillapproximately
approximately
10–20km
10–20 kmfrom
fromthethemouth.
mouth. The
The floating
floating section of the bridge
bridge occupies
occupies ~85%
~85%of ofthe
thewidth
widthofofHood
Hood
Canal, with
Canal, with a fixed opening
openingatateither
eitherend
endforfor
small vessel
small traffic.
vessel TheThe
traffic. bridge has ahas
bridge design draft draft
a design of 4.57of
m. As
4.57 m. part of the
As part Salish
of the SeaSea
Salish model
modeldevelopment
development process,
process, presence ofofthe
presence thefloating
floatingbridge
bridgewaswas
consideredbut
considered butnot
notincluded
included inin the
the model set-up under thethe assumption
assumptionthatthateffects
effectsofofthe
thebridge
bridgewere
were
likelyaalocal
likely local phenomenon
phenomenon and and unlikely to have a significant
significant impact
impact on
on the
thelarger-scale
larger-scaleSalish
SalishSea
Sea
model performance.
model performance.

Figure 1. (a) Oceanographic regions of Salish Sea including Northwest Straits, Puget Sound, and the
Figure 1. (a) Oceanographic regions of Salish Sea including Northwest Straits, Puget Sound, and the
inner sub-basins—Hood Canal, Whidbey Basin, Central Basin, and South Sound; (b) bathymetric
inner sub-basins—Hood Canal, Whidbey Basin, Central Basin, and South Sound; (b) bathymetric
profile of Hood Canal study area.
profile of Hood Canal study area.
J. Mar. Sci. Eng. 2018, 6, 119 3 of 22

However, recent research suggests that the bridge may, in fact, alter hydrodynamics and has the
potential to increase flushing time in Hood Canal. Circulation here exhibits classic fjord characteristics
of a shallow brackish layer at the surface over a deep long and narrow saltwater column that is
vulnerable to disruptions due to the presence of floating structures, which could constrict the mixing
and transport in the upper layers of the water column [7]. Recent studies also indicate that the bridge
is a barrier to fish passage. Slower migration times, higher mortality rates in the vicinity of the bridge
relative to other areas on their migration route, and unique behavior and mortality patterns at the
bridge suggest the bridge is impeding steelhead migration and increasing predation [8].
In this paper, we present an assessment of the near-field impact of the floating bridge on the
tidal hydrodynamics in the Hood Canal fjord environment as a component of the Hood Canal Bridge
Environmental Impact Assessment study [9]. In particular, we focus on quantifying the spatial
extent of the change in hydrodynamic parameters, such as currents, salinity, and temperature, in the
vicinity of the bridge relative to ambient conditions. This near-field region of deviation from ambient
background is the zone of influence (ZOI), defined arbitrarily as the distance from the bridge where
relative difference induced by the structures reduces to <10% of the maximum deviation. This was
accomplished through an application of the Salish Sea model [4] with the HCB embedded in high
resolution. We present a summary of field data collected at the bridge, set-up and calibration of the
Salish Sea model with a floating bridge module, and application of the model to the development of
quantitative estimates of ZOI dimensions. These results inform companion studies addressing the
impacts on the swimming behavior and observed mortality of out-migrating juvenile steelhead at
HCB. Although this study discusses site-specific efforts, the results and the concern about potential
ecological impacts are applicable to all floating bridges worldwide.

2. Materials and Methods

2.1. Oceanographic Data Collection

Oceanographic data collection was planned with two objectives: (1) field confirmation of the
hypothesis that the HCB obstructs and alters near-field surface currents; (2) provide near-bridge data
for the calibration of hydrodynamic models. Near field, for the purpose of this scope, is defined as the
region where the influence of the bridge on current, salinity, and temperature variables is noticeable
relative to the ambient (far field). Based on a prior modeling effort [7], the ZOI was expected to
vary from as small as one to two bridge widths (18–36 m) normal to the direction of flow, to a much
larger region covering several Hood Canal channel widths (2.4–7.2 km), for variables such as currents,
temperature, and salinity.
In situ current and conductivity, temperature, and depth (hereafter CTD) measurements were
collected over a four-week period, from late April to early June 2017, at three locations near
HCB [10]. Figure 2 shows the locations of the stations (a) immediately below the floating bridge span,
(b) approximately 500 m upstream (south) of the bridge, and (c) approximately 500 m downstream
(north) of the bridge. The bridge-mounted current meter (a) was attached to the floating section of the
bridge. Attachment of the acoustic Doppler current profiler (ADCP) to this platform provided a profile
of the water currents directly below the bridge. In addition to the ADCP, the bridge-mounted system
also included a single-point Aquadopp current meter attached approximately 1 m below the hull of the
bridge to collect in situ current data as close to the structure as feasible. The upstream and downstream
deployment locations were constrained by the fact that they needed to be bottom-mounted (outside of
marine traffic in the surface waters) and away from HCB mooring lines. Also, the maximum water
depth at the bridge central span was ~101 m and posed a challenge for the upward-looking ADCP
instrument’s ability to penetrate the surface layers. They were, therefore, deployed at a depth of
approximately 50 m and as far from the shoreline as possible. Continuous CTD measurements were
obtained from the bridge and bottom-mounted stations along with CTD casts during deployment
and recovery.
J. Mar. Sci. Eng. 2018, 6, 119 4 of 22
J. Mar. Sci. Eng. 2018, 6, x 4 of 22

Figure 2.2.Hood
Hood Canal
Canal Bridge
Bridge (HCB)(HCB) study
study area area showing
showing thebridge
the floating floating bridge
and the 2017 and the 2017
oceanographic
oceanographic data collection
data collection program stationprogram
locationsstation locations
relative relative to the bridge.
to the bridge.

2.2. The Salish

2.2. The Salish Sea
Sea Model
Model Set-Up
Set-Up
The Salish
The Salish Sea Sea model
model is is an
an externally
externally coupled
coupled hydrodynamic
hydrodynamic and and biogeochemical
biogeochemical model model
developed by the Pacific Northwest National Laboratory in collaboration
developed by the Pacific Northwest National Laboratory in collaboration with Washington with Washington State State
Department of
Department of Ecology
Ecology to to support
support coastal
coastal estuarine
estuarine research,
research, restoration
restoration planning,
planning, water-quality
water-quality
management,
management, and andclimate-change
climate-change response assessments
response in the in
assessments region
the [2–4]
region The[2–4]
modelThewas model
constructedwas
using the unstructured grid Finite-Volume Community Ocean
constructed using the unstructured grid Finite-Volume Community Ocean Model (FVCOM) Model (FVCOM) [11] version
[11]
2.7 framework and integrated-compartment model biogeochemical
version 2.7 framework and integrated-compartment model biogeochemical water-quality kinetics water-quality kinetics [12,13].
To facilitate
[12,13]. enhancedenhanced
To facilitate exchangeexchange
with the with PacifictheOcean,
Pacificthe Salishthe
Ocean, SeaSalish
modelSea gridmodel
was expanded
grid was
to include to
expanded coastal
include waters
coastalaround
watersVancouver Island and
around Vancouver the continental
Island shelf from
and the continental Canada’s
shelf Queen
from Canada’s
Charlotte Strait to Oregon’s Waldport, south of Yaquina Bay [14]. The
Queen Charlotte Strait to Oregon’s Waldport, south of Yaquina Bay [14]. The model kinetics were model kinetics were also
improved through the addition of sediment diagenesis and carbonate chemistry
also improved through the addition of sediment diagenesis and carbonate chemistry [15], and it is [15], and it is capable
of reproducing
capable the observed
of reproducing thebiogeochemical conditions ofconditions
observed biogeochemical the Salish Sea (salinity
of the Salish(S),Sea
temperature
(salinity (S),(T),
dissolved oxygen (DO), nitrate, algal biomass, and pH), including near-bed
temperature (T), dissolved oxygen (DO), nitrate, algal biomass, and pH), including near-bed hypoxia hypoxia in key locations
such
in keyaslocations
the Lynch Cove
such as region
the Lynchof Hood
CoveCanal,
regionPenn Cove,
of Hood and East
Canal, PennSound
Cove, [4].
and East Sound [4].
Figure 3a shows the Salish Sea model domain and grid
Figure 3a shows the Salish Sea model domain and grid with 16,012 nodes and with 16,012 nodes and 25,019
25,019 triangular
triangular
elements. The vertical configuration of the model uses 10 sigma-stretched
elements. The vertical configuration of the model uses 10 sigma-stretched layers distributed using layers distributed usinga
a powerlaw
power lawfunction
functionwithwithan anexponent
exponentP-SigmaP-Sigmaofof1.5,1.5,which
whichprovides
provides moremore layer
layer density
density nearnear the
the
surface.
surface. TheThemodel
modelwas wassetset
upupforfor
thetheyearyear
2017 2017 covering
covering the near-field
the near-field data data collection
collection period period of
of April
April through June. The model is loaded with daily values of freshwater
through June. The model is loaded with daily values of freshwater inflow from a total of 23 major inflow from a total of 23
major gauged
gauged rivers, rivers, 46 ungauged
46 ungauged streamsstreams
for which forflows
whichwereflows were estimated
estimated through hydrological
through hydrological analysis,
analysis, and 100 wastewater flows [16]. The model is forced with wind
and 100 wastewater flows [16]. The model is forced with wind and heat flux at the water surface. and heat flux at the water
surface. Meteorological
Meteorological inputs wereinputs were obtained
obtained from Weather
from Weather Forecasting
Forecasting ResearchResearch
ModelModel reanalysis
reanalysis data
data generated by the University of Washington. Tidal forcing at the open
generated by the University of Washington. Tidal forcing at the open boundary was based on tidal boundary was based on
tidal constituents (S2, M2, N2, K2, K1, P1, O1, Q1, M4, and M6) derived from
constituents (S2, M2, N2, K2, K1, P1, O1, Q1, M4, and M6) derived from the Eastern North Pacific or the Eastern North Pacific
or ENPAC
ENPAC model
model [17].
[17]. Temperature
Temperature andandsalinity
salinityprofiles
profilesatatthe
theboundary
boundarywere wereextracted
extracted from
from World
World
Ocean Atlas 2013 version 2 [18,19] climatological fields of in situ temperature and salinity at a 1 ◦ grid
Ocean Atlas 2013 version 2 [18,19] climatological fields of in situ temperature and salinity at a 1° grid
interpolated to the model boundary nodes. Spin up for the hydrodynamic model was conducted
through a one-year run initiated from stationary conditions (zero initial velocity and water surface
J. Mar. Sci. Eng. 2018, 6, 119 5 of 22

interpolated to2018,
J. Mar. Sci. Eng. the6,model
x boundary nodes. Spin up for the hydrodynamic model was conducted 5 of 22
through a one-year run initiated from stationary conditions (zero initial velocity and water surface
elevation)
elevation) with
with uniformTT(7(7◦ C)
uniform °C)and
andSS(32(32practical
practical salinity
salinity units
units (PSU)).
(PSU)).The
Thesimulation
simulation waswasthen
then
repeated using year-end results as the restart initial
repeated using year-end results as the restart initial condition. condition.
AsAs a preliminary
a preliminary step,
step, Salish
Salish SeaSea modelvalidation
model validationoveroverthe
theentire
entiredomain
domainwas wasfirst
firstconducted
conducted to
to ensure that modification of the model grid and layering scheme (described
ensure that modification of the model grid and layering scheme (described in the following in the following section)
section)
did not alter the overall model performance. This was done using year-long monitoring data available
did not alter the overall model performance. This was done using year-long monitoring data available
from a prior year (2014). The error statistics of water-surface elevation, S, and T were computed at
from a prior year (2014). The error statistics of water-surface elevation, S, and T were computed at nine
nine tide stations and 24 water-quality stations maintained by Washington State Department of
tide stations and 24 water-quality stations maintained by Washington State Department of Ecology and
Ecology and shown in Figure 3b. Overall calibration results for T and S at all stations were found to
shown in Figure 3b. Overall calibration results for T and S at all stations were found to be reasonable.
be reasonable. Relative water-surface elevation errors were less than 10% at all stations within Puget
Relative water-surface elevation errors were less than 10% at all stations within Puget Sound in United
Sound in United States (US) waters. Domain-wide T root-mean-square error (RMSE) was 0.94 °C with
◦ with a bias of −0.03 ◦ C.
States (US)
a bias waters.
of −0.03 °C.Domain-wide
The domain-wide T root-mean-square
S error was 0.96 error (RMSE)
PSU with was
a bias of 0.94
−0.24 CPSU. Model skill scores
Thewere
domain-wide
also high for S error
T andwas 0.96 Willmott
S with PSU withskill
a bias of −
score 0.24 PSU.
(WSS) Modelofskill
[20] values 0.96scores wererespectively
and 0.80, also high for
T and
(seeSTable
with 1).Willmott skill score (WSS) [20] values of 0.96 and 0.80, respectively (see Table 1).

Figure 3. (a)
Figure Salish
3. (a) Sea
Salish Seamodel
modeldomain
domainand
andfinite-volume grid; (b)
finite-volume grid; (b)monthly
monthlywater-quality
water-qualitymonitoring
monitoring
stations from Washington State Department of Ecology.
stations from Washington State Department of Ecology.

Table 1. Salish Sea model overall error statistics and skill score for temperature and salinity calibration
Table 1. Salish Sea model overall error statistics and skill score for temperature and salinity calibration
data for the year 2014 from monthly monitoring profiles.
data for the year 2014 from monthly monitoring profiles.
Mean
MeanError
Error(Bias)
(Bias) RMSE
RMSE WSS
WSS
Max Min
Max Min Ave
Ave Max
Max Min
Min Ave
Ave MaxMax Min Min Ave Ave
T (◦TC)(°C) 0.97
0.97 −1.77 −0.03
−1.77 −0.03 2.29
2.29 0.40
0.40 0.94
0.94 0.99
0.99 0.78
0.78 0.94
0.94
S (PSU) 0.09 −0.93 −0.24 2.27 0.39 0.96 0.90 0.60 0.80
S (PSU) 0.09 −0.93 −0.24 2.27 0.39 0.96 0.90 0.60 0.80
RMSE = root-mean-square error; WSS = Willmott skill score; T = temperature; S = salinity; PSU =
RMSE = root-mean-square error; WSS = Willmott skill score; T = temperature; S = salinity; PSU = practical salinity
practical
unit; salinity unit;
Max = maximum valueMax = maximum
of site-specific errorvalue of among
statistic site-specific error
21 Puget statistic
Sound among
stations; Min =21minimum
Puget Sound
value of
stations; error
site-specific Min statistic
= minimumamongvalue of site-specific
21 Puget errorAve
Sound stations; statistic among
= global error 21 Pugetconsidering
statistic Sound stations; Ave
data and =
model
results from
global all stations.
error statistic considering data and model results from all stations.

2.3.2.3.
Hood Canal
Hood Bridge
Canal BridgeModule
ModuleImplementation
Implementation

TheThe FVCOMframework
FVCOM frameworkusedused in
in the
the Salish
Salish Sea model
model isis aa three-dimensional,
three-dimensional,free-surface,
free-surface,
terrain-followingsolver
terrain-following solver for
for the
the primitive
primitive form
formofofNavier–Stokes
Navier–Stokes equations. The The
equations. model uses an
model uses
unstructured grid in the horizontal dimension and sigma-stretched coordinate system in the vertical
dimension. The floating bridge in this framework may be best approximated by a rectangular barge-
J. Mar. Sci. Eng. 2018, 6, 119 6 of 22

an unstructured grid in the horizontal dimension and sigma-stretched coordinate system in the
vertical dimension. The floating bridge in this framework may be best approximated by a rectangular
barge-like pontoon (see Figure 4 for HCB profile and cross-section details). However, representation
of the rectilinear shape of the pontoons in this discretized framework requires special attention.
While the width and the horizontal dimensions may be represented in a straightforward manner by
suitable refinement and arrangement of the triangular elements, representing the uniform draft in
a terrain-following sigma coordinate system poses a particular challenge. The dimensions of the bridge
were specified as 1992 m long and 18.3 m wide, with a draft of 4.57 m. Figure 5 shows a close-up of
the model grid refinement for the HCB region of the domain. The following three techniques were
explored:

(a) Implementation of a velocity block: This approach is identical to that described by Khangaonkar
and Wang (2013) [7] where the impermeable surface block was incorporated into FVCOM
with modification of both external and internal modes of the solver. For the baroclinic internal
mode, the horizontal velocities at the selected cells and surface layers were always specified as
zero such that no horizontal flow was allowed to pass through. During the barotropic external
mode calculations, the cross-sectional water column depth at selected cells occupied by the
block was adjusted to a new reduced value by subtracting the blocked layer thickness from the
total water depth. This modification accommodates the presence of the rigid structure but is
an approximation, as non-hydrostatic components of the pressure term, which are likely to be
strong in the near field, are neglected. Effects of the bridge on momentum terms are addressed,
but are done so as an indirect effect of setting the surface boundary to zero velocity without
affecting the pressure term.
(b) Implementation of momentum sink at the bridge using form drag: In this approach, the cells occupying
the bridge are populated with hypothetical cylinders similar to a densely packed kelp farm.
The drag from the cylinders set to sufficiently high value results in blockage of nearly ~95% of
surface currents. Although this represents a leaking bridge, the implementation allows effects on
continuity, as well as momentum, terms of the governing equations. The implementation of form
drag from suspended cylinders in the water column was described by Wang et al. (2013) [21].
This method also requires local modification of the bathymetry to a representative average depth
under the bridge for representation of the rectangular shape of the bridge pontoons.
(c) Free surface pressure modification with a bottom drag: This method relies on modification of the free
surface pressure boundary condition; an increase in pressure equivalent of 4.57 m of head results
in a model response of 4.57 m depression of the free surface. This method is an improvement over
(a) and (b) in that bathymetry is unaltered. In addition to modifying the free surface, the method
also employs drag formulation for the layer immediately under the bridge. This results in flow
passing the bridge under modified pressure with suitable reduction in velocity induced by the
form drag.

The efficacy of each method was tested by examination of predicted velocity, temperature,
and salinity profiles relative to the data collected. Results showed that all methods provided similar
performance in terms of near-field impacts to the flow field. For simplicity, the continuity/velocity
block method (a) was retained for the remainder of the analysis.
J. Mar. Sci. Eng. 2018, 6, 119 7 of 22
J. Mar. Sci. Eng. 2018, 6, x 7 of 22

Figure 4. HCB layout with pontoon cross-section and east span details. Detail
Detail A
A is
is a sectional view
across the width of the bridge.
J. Mar. Sci. Eng. 2018, 6, x 8 of 22
J. Mar. Sci. Eng. 2018, 6, 119 8 of 22
J. Mar. Sci. Eng. 2018, 6, x 8 of 22

Figure 5. Salish Sea model grid with refinement in the region to facilitate incorporation of the bridge
block
Figureelements.
Figure 5.5.Salish Themodel
SalishSea
Sea grid size
model gridis
grid refined
with
with such that
refinement
refinement in cell
inthe
the centers
region
region to are separated
tofacilitate
facilitate by the bridge
incorporation
incorporation of
ofthe width
thebridge
bridge
distance
block of 18.3The
blockelements.
elements. m. The
grid
The bridge
sizesize
grid draft
is refinedis such
represented
is refined thatby
that cell
such twocenters
centers
cell layers whose
are separated
are combined by thickness
by the bridge
separated thewidth
bridge is width
equal
distance
to
of 4.57
18.3 m.
m. The bridge draft is represented by two layers whose combined thickness is
distance of 18.3 m. The bridge draft is represented by two layers whose combined thickness is equal equal to 4.57 m.
to 4.57 m.
2.4.
2.4. Near-Field
Near-Field Model
Model Validation
Validation
2.4. Near-Field
The Model
The near-field
near-field Validation
model
model validation
validation consisted
consisted of of comparison
comparison of of predicted
predicted water
water surface
surface elevations,
elevations,
currents,
currents, salinity,
The salinity,
near-field and
and temperature
temperature
model validationresults
results with data
with data
consisted measurements
measurements
of comparison during the
during the
of predicted 2017
2017
water field program.
field program.
surface elevations,
Water surface
surface elevations:
Watersalinity,
currents, elevations: Water
Water surface
and temperature resultselevation
surface with datameasurements
elevation measurementswere
measurements were collected
duringcollected at
the 2017at aafield
station
station located
located
program.
on
on the
the east
east bank
bank of ofHood
Hood Canal.
Canal.Figure
Figure 6 shows
6 showsa comparison
a comparison of measured
of measured data
Water surface elevations: Water surface elevation measurements were collected at a station located and
data water
and surface
water level
surface
simulated
level
on the by
simulated the
east bankby model.
ofthe
Hood Absolute
model. mean
Absolute
Canal. error
Figure mean
6 shows (AME)
error or bias at
(AME) or bias
a comparison this station was
at this station
of measured 0.013 m,
was water
data and and
0.013 m,RMSE
and
surface
was
RMSE 0.42 m,
was which
0.42 m, is an
which error
is of
an 8%
error relative
of 8% to the
relative tidal
to range
the tidalof 5.34
range m
of
level simulated by the model. Absolute mean error (AME) or bias at this station was 0.013 m, and(Table
5.34 m 2).
(Table 2).
RMSE Salinity
Salinity and
andm,
was 0.42 temperature:
temperature:
which is an In addition
In error
addition
of 8%to relative
to velocityto
velocity profiles,
profiles,
the tidal North
North ADCP,
of 5.34Bridge
rangeADCP, Bridge
m (TableADCP,
ADCP, 2). andand South
South
ADCP
ADCP stations
stations
Salinity also
andalso provided salinity
providedInsalinity
temperature: addition and
and temperature
temperature
to velocity data.
data.
profiles, Temperature
Temperature
North ADCP, Bridgeand salinity
and salinity
ADCP, data data were
were
and South
measured
measured at
at aa depth
depth ofof11m m below
below the
thebridge
bridge hull
hull(5.57
(5.57m depth)
m depth) at the
at
ADCP stations also provided salinity and temperature data. Temperature and salinity data were Bridge
the Bridge ADCP
ADCP station, while
station, whileat
the
at North
the North
measured ADCP
at ADCP and
a depthand South
of 1South ADCP
ADCP
m below stations,
thestations, these data
these(5.57
bridge hull were
datamwere collected
depth)collected at
at the at a depth
a depth
Bridge of ~50
of ~50
ADCP m near
m near
station, the
the
while
seabed.
seabed. Figure
Figure
at the North 77 shows
ADCP shows the time-series
the
and South time-series
ADCP stations,comparison
comparison
these data of observed
of observed data
data and
were collected and simulated
simulated
at a depth of ~50 salinity
m nearand
salinity and
the
temperature
temperature at
at the
the three
three ADCP
ADCP locations.
locations. The
The salinity
salinity variation
variation at
at
seabed. Figure 7 shows the time-series comparison of observed data and simulated salinity and the
the three
three ADCP
ADCP stations
stations was
was
successfully
temperaturereproduced
successfully reproduced
at the three ininthe
ADCPmodel
the model simulations.
simulations.
locations. Predicted
Predicted
The salinity temperature
temperature
variation at thevariations were were
variations
three ADCP also inalso
stations goodin
was
agreement
good with
agreement observed
with data.
observed data.
successfully reproduced in the model simulations. Predicted temperature variations were also in
good agreement with observed data.

Figure 6. Time
Figure Timeseries
seriesofofmodel
modelresult,
result,NOAA
NOAA harmonic
harmonic (Xtide) prediction,
(Xtide) andand
prediction, observed datadata
observed for sea-
for
water elevation.
sea-water Time
elevation. is
Time shown
is as
shown Julian
as days
Julian in
days 2017.
in 2017.
Figure 6. Time series of model result, NOAA harmonic (Xtide) prediction, and observed data for sea-
water elevation. Time is shown as Julian days in 2017.
J. Mar. Sci. Eng. 2018, 6, 119 9 of 22
J. Mar. Sci. Eng. 2018, 6, x 9 of 22

Figure7.7. Time
Figure Time series
series of
of model
model results
results and observed
observed data
data for
for salinity
salinity (left
(left panel)
panel)and
andtemperature
temperature
(rightpanel)
(right panel)from
from3131May
Mayandand 11 June 2017. Plots (a,d)
(a,d) are
areaacomparison
comparisonfor forthe
theNorth
Northacoustic
acousticDoppler
Doppler
current profiler (ADCP) station;
current profiler (ADCP) station; plots plots (b,e) are for the Bridge
Bridge ADCP station; and plots (c,f) arefor
ADCP station; and plots (c,f) are forthe
the
SouthADCP
South ADCPstation.
station.

Table2.2.Salish
Table SeaSea
Salish model calibration
model error
calibration statistics
error and skill
statistics and score
skill for water
score for elevation. AME = AME
water elevation. absolute
=
mean error.
absolute mean error.
AME
AME RMSE
RMSE WSS
WSS
Elevation(m)
Elevation (m) 0.013
0.013 0.420
0.420 0.960
0.960

Theerror
The errorstatistics
statistics(AME
(AME and
and RMSE)
RMSE) between
between model
model predictions
predictions andand
fieldfield observations
observations are
are listed
listed in Table 3. The AME and RMSE for salinity for most stations were less than 1 PSU except for
in Table 3. The AME and RMSE for salinity for most stations were less than 1 PSU except for the
the Bridge ADCP station. The model predicted higher salinity compared to observed data at Julian
Bridge ADCP station. The model predicted higher salinity compared to observed data at Julian day
day 151 when the salinity dropped to 22 PSU. The temperature AME and RMSE were <1 °C for all
151 when the salinity dropped to 22 PSU. The temperature AME and RMSE were <1 ◦ C for all stations,
stations, demonstrating a good match of model predictions with temperature data near the bridge.
demonstrating a good match of model predictions with temperature data near the bridge.
Table 3. Salish Sea model validation error statistics and skill score for temperature and salinity at
Table 3. Salish Sea model validation error statistics and skill score for temperature and salinity
different locations near Hood Canal Bridge (HCB). ADCP = acoustic Doppler current profiler; CTD =
at different locations near Hood Canal Bridge (HCB). ADCP = acoustic Doppler current profiler;
conductivity, temperature, and depth.
CTD = conductivity, temperature, and depth.
Temperature Salinity
Station Temperature Salinity
Station AME RMSE WSS AME RMSE WSS
North ADCP AME0.22 RMSE
0.56 WSS
0.74 AME 0.57
0.21 RMSE 0.30WSS
North ADCP
South ADCP 0.22
0.08 0.56
0.40 0.74
0.85 0.21 0.360.57 0.41 0.30
0.12
South ADCP 0.08
Bridge ADCP 0.41 0.40
0.79 0.85
0.90 0.12 1.290.36 0.47 0.41
0.82
Bridge ADCP 0.41 0.79 0.90 0.82 1.29 0.47
North CTD 0.47
North CTD
0.47 0.91
0.91 0.80
0.80
0.33
0.33
0.860.86 0.68 0.68
South
South CTDCTD 0.500.50 0.80
0.80 0.90
0.90 0.65
0.65 1.031.03 0.72 0.72

A series of CTD casts were also conducted north and south of the bridge. These CTD casts were
A series of CTD casts were also conducted north and south of the bridge. These CTD casts were
collected at peak ebb and peak flood, as well as during slack water conditions. The observed salinity
collected at peak ebb and peak flood, as well as during slack water conditions. The observed salinity
and temperature profiles were then compared with simulated temperature and salinity at the same
and
location and time.profiles
temperature Figureswere then
8 and compared
9 show withofsimulated
examples temperature
comparisons and salinity
between observed and at the same
simulated
location and time. Figures 8 and 9 show examples of comparisons between observed and
salinity and temperature profiles at the north and south CTD cast locations, respectively. simulated
salinity and temperature profiles at the north and south CTD cast locations, respectively.
J.
J. Mar.
Mar. Sci.
Sci. Eng.
Eng. 2018,
2018, 6,
6, x119 10
10 of
of 22
22
J. Mar. Sci. Eng. 2018, 6, x 10 of 22

Figure 8. Comparisons of predicted and observed salinities at peak ebb and peak flood north of HCB
Figure 8. Comparisons of predicted and observed salinities at peak ebb and peak flood north of HCB
on 31 May 2017.
on 31
on 31 May
May 2017.
2017.

Figure 9. Comparisons of predicted and observed salinities at peak ebb and peak flood south of HCB
Figure 9. Comparisons of predicted and observed salinities at peak ebb and peak flood south of HCB
on 31 May 2017.
on 31 May 2017.
The error statistics of temperature and salinity at the north and south CTD locations are also
listedThe error 3.statistics of temperature and salinityhadat the north and south CTD locations
◦C ◦ C. For are also
in Table For temperature, both locations AME ~0.5 °C with
with RMSE
RMSE <0.8
<0.8 °C. salinity,
For salinity,
listed and
AME in Table
RMSE 3. were
For temperature,
0.65 and 1.03both PSU,locations had Although
respectively. AME ~0.5overall°C with RMSE
error <0.8 °C.
statistics forFor
thesalinity,
profiles
PSU, respectively.
AME and RMSE were 0.65 and 1.03 PSU, respectively. Although overall error statistics
were acceptable, examination of the results, especially for salinity, shows that the stratification in the for the profiles
were
upperacceptable,
10% of theexamination
water column of in
thethe
results,
modelespecially for salinity,
is not as strong as in shows that
the data. thecould
This stratification in the
be due to the
upper 10% of the water column in the model is not as strong as in bethe data. This could be due to the
limitation of the model confined to 10 sigma-layers. It could also be associated
associated with
with near-field
near-field mixing
limitation of the model confined to 10 sigma-layers. It could also be associated with near-field mixing
induced by the bridge module as incorporated using the cell velocity block option. The results were
induced
similar forbythethethree
bridge moduleused
methods as incorporated using the cell
for approximating velocity block option. The results were
approximating the the bridge
bridge block.
block.
similar for theCurrent
Currents: three methods
Current usedatfor
predictions
predictions theapproximating
at North
the NorthADCP ADCP the station,
bridge
station, block.
the South ADCPADCP
the South station, and theand
station, Bridge
the
ADCP Currents: Current predictions at the North ADCP station, the SouthwereADCP station,depths
and the
Bridge station
ADCP werestationcompared against field
were compared observations.
against The comparisons
field observations. The comparisons selected
wereatselected of
at
Bridge
4 m, 9 ADCP
m, 24 m,station
39 m, were
54 m, compared
and 69 m. against
To better field observations.
illustrate the The comparisons
comparison, velocities were
were selected at
decomposed
depths of 4 m, 9 m, 24 m, 39 m, 54 m, and 69 m. To better illustrate the comparison, velocities were
depths of 4 m, 9 m,
into longitudinal 24transversal
m, 39 m, 54 m, and 69 m. andTo better illustrate the comparison, velocities were
decomposed intoandlongitudinal components
and transversal are provided
components in Figure
and 10. The predicted
are provided in Figure phase of
10. The
decomposed
the velocity into
and longitudinal
the velocity and
magnitude transversal
associated components
with flooding and are
and provided
ebbing are in
in Figure
good 10. The
agreement
predicted phase of the velocity and the velocity magnitude associated with flooding and ebbing are
predicted
with phase
the field of the velocity
observations. and theshows
velocity magnitude of associated with
andflooding and ebbing are
in good agreement with theFigure
field 10 a comparison
observations. Figure 10 shows predicted observed
a comparison velocities
of predicted for
and
in good
observed agreement
different depths
velocities with
at the the
for Bridge field
different ADCP observations.
depthsStation.
at the At Figure
a depth
Bridge ADCP10 shows
of 9Station. a comparison
m, the ADCP data of
At a depth of
bin predicted
9 nearest toand
m, the ADCP the
observed
surface, velocities for different depths at the Bridge ADCP Station.velocities.
At a depth of 9 m, the ADCP
data bin the effect
nearest tooftheHCB is reflected
surface, in both
the effect of HCBpredicted andin
is reflected observed
both predicted and observed velocities.
data bin nearest to the surface, the effect of HCB is reflected in both predicted and observed velocities.
J. Mar. Sci. Eng. 2018, 6, 119 11 of 22
J. Mar. Sci. Eng. 2018, 6, x 11 of 22

10. Comparisons
Figure 10. Comparisons of predicted
predicted (blue)
(blue) and observed
observed (red) velocities
velocities for different depths at the
Bridge ADCP station (example).

In addition
In addition totoADCP
ADCPmeasurements,
measurements,ininanan attempt
attempt to to
getget
as close to the
as close bridge,
to the pontoon,
bridge, and and
pontoon, hull
as possible,
hull a single-point
as possible, measurement
a single-point measurement was was
conducted
conductedusingusing
an Aquadopp
an Aquadopp instrument, located
instrument, 1m
located
below HCB. Figure 11 shows a comparison between predicted currents and
1 m below HCB. Figure 11 shows a comparison between predicted currents and observed data. The observed data. The model
appearsappears
model to be overestimating the velocity
to be overestimating during peak
the velocity ebb.
during Thisebb.
peak region
Thisimmediately below thebelow
region immediately bridge is
the
known
bridge istoknown
generatetoeddies
generateandeddies
shed vortices
and shed asvortices
the ebbing water
as the is forced
ebbing water under the bridge.
is forced Thebridge.
under the model
is unable
The modeltoisresolve
unable these fine-scale
to resolve these eddies, which
fine-scale is one
eddies, whichexplanation for the error.
is one explanation Due
for the to the
error. Duefact
to
that the that
the fact modelthecurrently does notdoes
model currently include skin friction
not include at the bridge,
skin friction the slipthe
at the bridge, velocity in the model
slip velocity in the
immediately
model below the
immediately bridge
below is, therefore,
the bridge higher than
is, therefore, higher thethan
data.the data.

Figure 11.
Figure Comparisons of
11. Comparisons of predicted
predicted and
and observed
observed velocities
velocities just
just below
below HCB.
HCB.

The
The corresponding
correspondingerror statistics
error for for
statistics current comparisons
current are summarized
comparisons in Table
are summarized in4.Table
All stations
4. All
had AME <0.03 m/s. The RMSE for all stations was <0.21 m/s, except for the Aquadopp
stations had AME <0.03 m/s. The RMSE for all stations was <0.21 m/s, except for the Aquadopp single-
single-point measurement, where a higher RMSE of 0.33 m/s was noted. It is noted that these
point measurement, where a higher RMSE of 0.33 m/s was noted. It is noted that these errors are of
errors are ofcomparable
magnitude magnitudeto comparable to the
the sustained sustained
swimming swimming
velocity velocityJuvenile
of a 120-mm of a 120-mm
CohoJuvenile Coho
Salmon (~0.64
Salmon (~0.64 m/s) and the maximum swimming velocity for a 60–63-mm Juvenile Chum Salmon
m/s) and the maximum swimming velocity for a 60–63-mm Juvenile Chum Salmon (~0.14 m/s)
(~0.14 m/s) [22,23]. The WSS for the Aquadropp station was low (0.36) compared to other stations
[22,23]. The WSS for the Aquadropp station was low (0.36) compared to other stations due to the
due to the complexity of the measurement. However, the WSS for the North ADCP, the South ADCP,
complexity of the measurement. However, the WSS for the North ADCP, the South ADCP, and the
and the Bridge ADCP was >0.89, suggesting that the model calibration results were of sufficient quality
Bridge ADCP was >0.89, suggesting that the model calibration results were of sufficient quality for
for
useuse in ZOI
in ZOI calculations.
calculations.
J. Mar. Sci. Eng. 2018, 6, 119 12 of 22
J. Mar. Sci. Eng. 2018, 6, x 12 of 22

Table4.4.Salish
Table SalishSea
Seamodel
modelcalibration
calibrationerror
errorstatistics
statisticsand
andskill
skillscore
scorefor
forvelocities.
velocities.

Station
Station AME
AME RMSE
RMSE WSS
WSS
North
North ADCPADCP0.000.00 0.14
0.14 0.94
0.94
South ADCP
South ADCP 0.020.02 0.21
0.21 0.89
0.89
Bridge ADCP
Bridge ADCP0.030.03 0.21
0.21 0.92
0.92
Aquadopp 0.02 0.33 0.36
Aquadopp 0.02 0.33 0.36

To
Toassess
assessthetheeffect
effectofofHCB
HCBon onnear-field
near-fieldcurrents,
currents,we
weexamined
examinedthe thedata
dataduring
duringpeak
peakebb
ebband
and
flood periods for every tidal cycle at the North, South, and Bridge ADCP stations.
flood periods for every tidal cycle at the North, South, and Bridge ADCP stations. The expectation The expectation
was
wasthat
thatthetheeffect of of
effect thethe
bridge on on
bridge currents would
currents wouldbe at
bea at
maximum
a maximum during peakpeak
during currents. Figure
currents. 12a
Figure
shows the time series of depth-averaged longitudinal velocity at the bridge. Red triangles
12a shows the time series of depth-averaged longitudinal velocity at the bridge. Red triangles and and yellow
triangles
yellow indicate
trianglesselected
indicatepeak ebb and
selected flood
peak times,
ebb andrespectively,
flood times, at which velocityatinformation
respectively, was
which velocity
extracted in each
information wastidal cycle. in
extracted The averages
each of predicted
tidal cycle. peak velocity
The averages profiles
of predicted peakduring flood
velocity and ebb
profiles are
during
shown in Figure 12b–g along with measured data.
flood and ebb are shown in Figure 12b–g along with measured data.

Figure 12.(a)
Figure12. (a)Time
Timeseries
seriesofofdepth-averaged
depth-averagedvelocity
velocityatatbridge;
bridge;(b–g)
(b–g)comparisons
comparisonsofofpredicted
predictedand
and
observed
observedaverage
averagevelocity
velocityprofiles
profilesatatthe
theSouth,
South,Bridge,
Bridge,andandNorth
NorthADCP
ADCPstations
stationsduring
duringmaximum
maximum
flood
floodand
andebb
ebb tide Graythin
tide periods. Gray thinlines
lines represent
represent daily
daily velocity
velocity profiles
profiles during
during peakpeak ebb flood
ebb and and
flood periods.
periods.
J. Mar. Sci. Eng. 2018, 6, 119 13 of 22
J. Mar. Sci. Eng. 2018, 6, x 13 of 22

Positive velocity
Positive velocity represents
represents ebbebbor oroutflowing
outflowingcurrent
currentdirection
directiondue
duenorth
northtoward
toward thethe
mouth
mouth of
Hood
of Hood Canal.
Canal. Negative
Negative velocity
velocityindicates
indicates flood
floodtidal
tidalcurrents
currentsdue
duesouth
southininthe
the opposite
opposite direction.
direction.
Figure 12b–g show that the model is able to match the peak observed average
Figure 12b–g show that the model is able to match the peak observed average velocity well at velocity well at all
all
stations during ebb and flood. The velocity reduction due to HCB is reproduced
stations during ebb and flood. The velocity reduction due to HCB is reproduced well by the model, well by the model,
as shown
as shown in inFigure
Figure12c,f.
12c,f.This
Thisconfirms
confirmsthat that the
the implementation
implementation of the
of the bridge
bridge module
module in Salish
in the the Salish
Sea
Sea model is successful in reproducing the reduction in velocity near the bridge. Small
model is successful in reproducing the reduction in velocity near the bridge. Small discrepancies were discrepancies
were found
found at all stations
at all stations that were that were attributed
attributed to approximations
to approximations associated associated with of
with smoothing smoothing
bathymetry of
bathymetry
and and grid resolution.
grid resolution. Similarly, representation
Similarly, representation of HCB in of HCB in a sigma-coordinate
a sigma-coordinate framework framework
required
required flattening of the depths immediately below the bridge, resulting in
flattening of the depths immediately below the bridge, resulting in slightly higher velocitiesslightly higher velocities
in the
in the column.
water water column.

3. Results

3.1. Zone of Impact—Surface Layer

To
To characterize
characterize the the bridge
bridge influence
influence onon the
the near-field
near-field environment,
environment, two two sensitivity test scenario
runs were performed. In test 1, the bridge block (1992 m long, 18.3 m wide, and 4.57 m deep) was
completely removed, representing the conditions without the bridge. In test 2, the central (movable)
span of the bridge with a length of 182 m was left left in
in the
the open
open position
position with with the
the bridge
bridge pontoons
present in place.
present in place. The central span is typically opened to allow ship and boat traffic to pass providing
a third
thirdopening
openingfor fortidal
tidaltransport,
transport,with thethe
with twotwopermanent
permanent openings present
openings at the
present ateast
the and
east west
and ends
west
of
endstheof
bridge. To characterize
the bridge. the ZOI,
To characterize thebaseline conditions
ZOI, baseline from the
conditions fromsummer of 2017of
the summer with the
2017 bridge
with the
in place were compared with the simulated results from the sensitivity tests
bridge in place were compared with the simulated results from the sensitivity tests (1) without the (1) without the bridge
and
bridge(2) and
with(2)thewith
central
the span open.
central spanComparisons
open. Comparisonswere conducted for predicted
were conducted currents,currents,
for predicted salinity,
and temperature,
salinity, and ZOI estimates
and temperature, were obtained
and ZOI estimates were through
obtainedquantifying the difference
through quantifying the relative to
difference
existing
relative toor existing
baselineor conditions.
baseline conditions.
Figure 13 shows contour and velocity vector plots of existing existing conditions
conditions with HCB during ebb
and flood. PeakPeakebbebband andflood
floodtimes
timeswere
wereselected
selectedduring
duringa atypical
typical spring
spring tidetide
onon27 27 April
April 2017,
2017, at
at 10:00
10:00 a.m.
a.m. and
and 6:00
6:00 p.m.,
p.m., respectively,for
respectively, forthis
thisexamination
examinationbased basedon on the
the expectation
expectation that maximum
impact of the bridge block, reducing the currents to zero, would be strongest during the spring tides
and during
during ebb
ebbandandflood
floodtimes.
times. Figure
Figure 13a,b
13a,b are are currents
currents withwith
HCBHCBpresent present
during during
the ebbtheand
ebbflood,
and
flood, respectively,
respectively, for thefor the surface
surface layer.layer. The surface
The surface layer layer occupies
occupies the upper
the upper 3% of3% theofwater
the water column.
column. The
The
ebb ebb currents
currents in the
in the surface
surface layers
layers areare stronger
stronger than
than thoseduring
those duringthetheflood.
flood.TheTheeffect
effect of
of the HCB
structure on currents is noticeable
noticeable at the
the bridge
bridge where
where velocity
velocity was
was set set to
to 00 m/s
m/s and a shadow of
reduced currents
currents exists
exists immediately
immediately behind
behind during
during both
both ebb
ebb and
and flood.
flood.

(a) (b)

Figure 13. Predicted horizontal velocity contour and vector plots for the baseline scenario with HCB
present for
present for (a)
(a) ebb
ebb and
and (b)
(b) flood
flood currents
currents in
in the
the surface
surfacelayer.
layer.(Typical
(Typicalspring
springtide
tideon
on27
27April
April2017).
2017).

Figure 14 shows the same simulation but without the bridge with a much larger region occupied
by currents in the >1 m/s bands.
J. Mar. Sci. Eng. 2018, 6, 119 14 of 22

Figure 14 shows the same simulation but without the bridge with a much larger region occupied
by
J. Mar. Sci. Eng.in
currents the6,>1
2018, x m/s bands. 14 of 22

(a) (b)

14. Predicted
Figure 14. Predicted horizontal
horizontal velocity
velocity contour
contour and
and vector
vector plots
plots for test 1 scenario without HCB
present for
present for (a)
(a) ebb
ebb and
and (b)
(b) flood
flood currents
currents in
in the
the surface
surface layer.
layer. (Typical
(Typicalspring
springtide
tideon
on27
27April
April2017).
2017).

Figure
Figure 1515shows
showssurface
surfacelayer currents
layer with
currents the central
with span span
the central open.open.
The current patterns
The current are similar
patterns are
to those at baseline conditions shown in Figure 13. Relative to the size (length) of the bridge
similar to those at baseline conditions shown in Figure 13. Relative to the size (length) of the bridge blocking
the surface
blocking thelayer, thelayer,
surface central
thespan represents
central an opening
span represents of ~9%. of
an opening The effect
~9%. Theofeffect
the opening on the
of the opening
currents is mostisvisible
on the currents in the form
most visible in theofform
the jet
ofbehind the bridge
the jet behind the during both ebb
bridge during andebb
both flood.
and flood.

(a) (b)

Figure 15. Predicted horizontal velocity contour and vector plots for test 2 scenario with HCB present
and middle span open for (a) ebb and (b) flood currents in the surface layer. (Typical spring
spring tide on
27 April 2017).

The ZOI
The ZOIisisbest characterized
best characterizedusing plots
using of difference
plots between
of difference simulated
between results from
simulated thefrom
results baseline
the
condition (with the bridge)
baseline condition (with the relative to tests
bridge) 1 and
relative to 2tests
to provide
1 and 2a to
direct assessment
provide of assessment
a direct the bridge impact.
of the
Figure 16a,b show
bridge impact. Figurea difference
16a,b show(reduction)
a differencein(reduction)
currents during peak during
in currents ebb and flood
peak ebbwith
and the
floodbridge
with
relative
the bridgeto relative
test 1 without
to test the bridge.the
1 without Figure
bridge.16c,d show
Figure a reduction
16c,d in currents
show a reduction in during
currentspeak ebbpeak
during and
flood relative to the condition with the bridge draw span open. The effect of the
ebb and flood relative to the condition with the bridge draw span open. The effect of the bridge on bridge on surface
currents is noticeable
surface currents over a distance
is noticeable of ~1–2of
over a distance Hood
~1–2Canal
Hoodwidths in either
Canal widths in direction (~2–5 km).
either direction (~2–5Effect
km).
of the bridge
Effect of the span
bridge opening on currents
span opening is limitedis(~0.25
on currents limitedHood Canal
(~0.25 Hoodwidth,
Canal<600 m), but
width, prominent
<600 m), but
during
prominentthe ebb tidethe
during withebbstronger surface
tide with currents.
stronger surface currents.
J. Mar. Sci. Eng. 2018, 6, 119 15 of 22
J. Mar. Sci. Eng. 2018, 6, x 15 of 22

(a) (b)

(c) (d)

16. (a,b)
Figure 16. (a,b)Velocity
Velocitydifference
differenceof of
baseline scenario
baseline withwith
scenario HCBHCB
relative to testto1 without
relative HCB during
test 1 without HCB
peak ebb
during andebb
peak flood;
and(c,d) velocity
flood; (c,d) difference of baseline
velocity difference ofscenario
baselinewith HCBwith
scenario relative
HCB to relative
test 2 with draw
to test 2
span open during ebb and flood. (Typical spring tide on 27 April 2017).
with draw span open during ebb and flood. (Typical spring tide on 27 April 2017).

These difference
These difference plots
plots in
in the
the plan
plan view provide aa qualitative
view provide estimate of
qualitative estimate of the
the size
size of
of the
the ZOI
ZOI forfor
currents
currents in
in the surface layer.
the surface layer. The spatial pattern
The spatial pattern and
and extents
extents were
were similar
similar in in nature
nature with
with the
the influence
influence
of HCB most prominently seen behind the bridge during peak ebb and
of HCB most prominently seen behind the bridge during peak ebb and flood conditions. In allflood conditions. In all cases,
cases,
the maximum difference of the baseline condition relative to the scenarios was
the maximum difference of the baseline condition relative to the scenarios was at the bridge itself. at the bridge itself.
In
In
thethe surface
surface layer,
layer, thethemaximum
maximumcurrent
currentdifference
difference was −0.88m/s,
was −0.88 m/s,the themaximum
maximumsalinity
salinity difference
difference
was +0.23 PSU,
PSU,andandthe themaximum
maximumtemperature
temperaturedifference
differencewas −0.49
was−0.49 ◦ C for the selected spring tide
was +0.23 °C for the selected spring tide on
on 27 April
27 April 2017.
2017. TheThe effect
effect of of
thethe bridge
bridge dissipateswith
dissipates withdistance
distancefrom
fromthe thestructure.
structure.Unlike
Unlikevelocity,
velocity,
where
where the
the effect
effect is
is directly
directly the
the result
result of
of blocked
blocked currents,
currents, the
the influence
influence on on salinity and temperature
salinity and temperature is is
more complex due to the presence of strong stratification and the increased mixing
more complex due to the presence of strong stratification and the increased mixing of surface layers of surface layers
at the
at the bridge. Vertical transects
bridge. Vertical analyzed in
transects analyzed in the
the next
next section
section provide
provide a a better
better understanding
understanding of of the
the
influence of the bridge on salinity and temperature.
influence of the bridge on salinity and temperature.

3.2. Zone of Impact—Vertical Transect

3.2. Zone of Impact—Vertical Transect
To further quantify the spatial extent of ZOI on current, salinity, and temperature variables in this
To further quantify the spatial extent of ZOI on current, salinity, and temperature variables in
environment with strong vertical stratification, simulation results were examined along a mid-channel
this environment with strong vertical stratification, simulation results were examined along a mid-
transect across HCB. The assessment was done using two transects: (a) Transect-a, approximately
channel transect across HCB. The assessment was done using two transects: (a) Transect-a,
2.5 km long (length scale of Hood Canal width) across the bridge to provide a detailed characterization
approximately 2.5 km long (length scale of Hood Canal width) across the bridge to provide a detailed
of near-field effects; and (b) Transect-b approximately 20 km long (distance from the bridge to the
characterization of near-field effects; and (b) Transect-b approximately 20 km long (distance from the
mouth of Hood Canal) to allow examination over larger distances over full tidal excursion (see
bridge to the mouth of Hood Canal) to allow examination over larger distances over full tidal
Figure 17). The results were analyzed as difference plots similar to the plan view maps in the previous
excursion (see Figure 17). The results were analyzed as difference plots similar to the plan view maps
section, during the selected typical spring tide of 27 April 2017, at peak ebb (147.42 Julian days (JD),
in the previous section, during the selected typical spring tide of 27 April 2017, at peak ebb (147.42
27 May 2017 10:00 a.m.) and peak flood times (147.75 JD, 27 May 2017 6:00 p.m.).
Julian days (JD), 27 May 2017 10:00 a.m.) and peak flood times (147.75 JD, 27 May 2017 6:00 p.m.).
J. Mar. Sci. Eng. 2018, 6, 119 16 of 22
J. Mar. Sci. Eng. 2018, 6, x 16 of 22
J. Mar. Sci. Eng. 2018, 6, x 16 of 22

(a) (b)
(a) (b)

Figure
Figure 17.17. Transectsfor
Transects forassessment
assessmentof
ofzone
zone of
of influence
influence (ZOI)
(ZOI) in
inaavertical
verticalplane:
plane:(a)(a)
short transect
short forfor
transect
Figure 17. effects;
near-field Transects
(b)for assessment
longer oftozone
transect of influence
examine the (ZOI)
effects in full
over a vertical
tidal plane: (a) short transect for
excursion.
near-field effects; (b) longer transect to examine the effects over full tidal excursion.
near-field effects; (b) longer transect to examine the effects over full tidal excursion.
Thenear-field
The near-field effects
effects of of HCB
HCBononthethe current,
current,salinity, and and
salinity, temperature
temperature in a vertical two-
in a vertical
The near-field
dimensional plane effects
near the ofZOI
HCBfor on
the the current,
example salinity,
spring tide and
periodtemperature
are shown inin aFigures
vertical18–20,
two-
two-dimensional plane near the ZOI for the example spring tide period are shown in Figures 18–20,
dimensional using
respectively, plane the
nearshorter
the ZOI for the example
Transect-a. spring
The distance
distance tide period
between −1.25 are shown in Figures 18–20,
respectively, using the shorter Transect-a. The between −1.25km kmandand0 0km km represents
represents thethe
respectively,
section south using
of the the shorter
bridge, while Transect-a.
that The
between 0distance
km and between
1.25 km −1.25 km
represents and
the 0 km represents
downstream the
section
section
section
south
south
ofofthe
the
bridge,
bridge,
while
while
that
that
between
between
00 km
km
and 1.25
and
1.25kmkmrepresents
representsthe thedownstream
downstream section
section
north
north of ofthethe bridge.As
bridge. Asininthe
thesurface
surfaceplan
plan view
view maps,
maps, aa reduction
reductionin incurrent
currentmagnitudes
magnitudes behind
behind thethe
north
bridge ofisthe
seenbridge.
during Asthe
in flood
the surface plan
and ebb. view 18a,b
Figure maps,show
a reduction
that theincurrents
current magnitudes
are behind
reduced behind the
the
bridge
bridgeis seen
is seenduring
during the
theflood
flood and
and ebb.
ebb. Figure
Figure 18a,b
18a,b show that
show thatthe
thecurrents
currentsare arereduced
reduced behind
behind thethe
bridge relative to the scenario without the bridge. The vertical transect shows that the effect extends
bridge relative
bridge relative to the scenario
to thetransect without
scenariolength
without the bridge.
the bridge. The
The vertical transect shows that the effect extends
beyond the 1.25-km on either side of thevertical
bridgetransect
and the shows
effect isthat the effect
noticeable extends
mostly in
beyond
beyond thethe1.25-km
1.25-km transect
transect length
length on
on either
either side
side of
of the bridge
the bridgeand andthetheeffect
effectisisnoticeable
noticeable mostly
mostly in in
the upper ~10–15 m of the water column. Figure 18c,d show a similar relative difference in currents
thethe
upper
upper ~10–15 mmofofthe water column. Figure 18c,d
18c,d show
showaasimilar
similarrelative
relativedifference
difference in currents
relative to ~10–15
the scenario thewith water column.
bridge draw Figure
span open. The effect is minor with a change of inup
currents
to 0.3
relative to
relative the
to scenario
the scenariowith bridge
with draw
bridge draw span open.
span The
open.
m/s seen at the draw span opening, extending <500 m from the bridge.The effect is
effect minor
is with
minor a
with change
a change of up
of to
up 0.30.3
to m/s
seen
m/satseen
the draw
at the span
drawopening,
span opening,extending <500<500
extending m from the bridge.
m from the bridge.

Figure 18. (a,b) Velocity difference of baseline scenario with HCB relative to test 1 without HCB
Figure
Figure
during 18.
18.peak(a,b)
(a,b) ebb Velocity
Velocity
and flood; difference
difference
(c,d) of of baseline
baseline
velocity scenario
scenario
difference ofwithwith
HCBHCB
baseline relative
relative
scenario to toHCB
test
with 1test 1 without
without HCB
relative HCB
during
to test 2
during
peak
withebb peak
andspan
draw ebb
flood;and flood;
(c,d)during
open (c,d)
velocity velocity
ebbdifference difference
and flood.of of
baseline
(Typical baseline
scenario
spring scenario
tide with
on 27HCB with HCB
Aprilrelative relative to test
2017). to test 2 with draw 2
with draw span open during ebb and flood. (Typical spring tide
span open during ebb and flood. (Typical spring tide on 27 April 2017). on 27 April 2017).

Figure 19. (a,b) Salinity difference of baseline scenario with HCB relative to test 1 without HCB during
Figure
Figure
peak 19.
19.
ebb (a,b)
(a,b)
and Salinity
Salinity
flood; difference
difference
(c,d) ofbaseline
velocity of baselineofscenario
difference baselinewith
scenario with HCB relative
HCBwith
scenario relative
HCB totorelative
test
test1 1without
to test HCB
without HCBduring
2 with during
draw
peak
peak ebb
ebb andand flood;
flood; (c,d)
(c,d) velocity
velocity difference
difference of
of baseline
baseline scenario
scenario
span open during ebb and flood. (Typical spring tide on 27 April 2017). with
with HCB
HCB relative
relativeto test
to 2
test with
2 draw
with draw
span
span open
open during
during ebbebband
andflood.
flood.(Typical
(Typicalspring
spring tide
tide on
on 27 April
April 2017).
2017).
J. Mar. Sci. Eng. 2018, 6, 119 17 of 22
J. Mar. Sci. Eng. 2018, 6, x 17 of 22

Figure 19a,b19a,bshow
showthe the effects of HCB
effects of HCB on salinity through
on salinity difference
through plots relative
difference to test 1to
plots relative without
test 1
the bridge, during ebb and flood, respectively. The surface layer salinity
without the bridge, during ebb and flood, respectively. The surface layer salinity immediately behind immediately behind the
bridge
the was was
bridge higher. This This
higher. was duewastodue thetoblocking effect of
the blocking the bridge
effect holdingholding
of the bridge the fresher
the (lower
freshersalinity)
(lower
water back
salinity) water during
backflood
during and ebb.and
flood Theebb.traversing surface layer
The traversing surface was forced
layer wasto flow to
forced underflowtheunderbridge,
the
resulting in reduced salinity in the deeper layers below the bridge. A similar
bridge, resulting in reduced salinity in the deeper layers below the bridge. A similar effect is seen effect is seen relative to
test 2 with
relative to the
testdraw
2 withspan theopen.
drawFigure
span 19c,d
open.show
Figure a small
19c,dincrease
show ainsmallsurface salinityinbehind
increase surface thesalinity
bridge
relative the
behind to the scenario
bridge with to
relative thethe
draw span open.
scenario with Plan view span
the draw plots are
open.consistent
Plan view withplots
velocity difference
are consistent
plots in that the
with velocity effect ofplots
difference test 2inwas
thatrestricted
the effect to of atest
small
2 wasregion near the
restricted to aspan
smallopening.
region near Salinity was
the span
lower in the
opening. layerswas
Salinity immediately
lower in the below the immediately
layers bridge due tobelow fresherthe water being
bridge dueforced to lower
to fresher layers
water by
being
the presence of the bridge relative to the condition with draw span open.
forced to lower layers by the presence of the bridge relative to the condition with draw span open.
The effect of HCB on temperature is similar to that predicted on salinity. salinity. Thermal stratification
surface heat flux results in water temperatures in the surface layers being being warmer
warmer thanthan lower
lower layers.
layers.
The presence of the bridge holds the surface layers back, resulting in cooler
presence of the bridge holds the surface layers back, resulting in cooler waters immediately waters immediately behind
it relative
behind to test 1to
it relative without the bridge
test 1 without theduring
bridgeebb and ebb
during flood,andasflood,
shownasinshown Figurein20a,b.
Figure The warmer
20a,b. The
water
warmer was forced
water wastoforced
traverse underneath
to traverse the bridge,
underneath resulting
the bridge, in warmer
resulting waters waters
in warmer in the deeper layers
in the deeper
behindbehind
layers the bridge. FigureFigure
the bridge. 20c,d 20c,d
showshow a similar effecteffect
a similar but onbutaon smaller
a smallerscale with
scale a small
with a smallincrease
increasein
surface
in temperature
surface temperature behind the bridge
behind relative relative
the bridge to the scenario
to thewith the draw
scenario withspantheopen.
drawTemperatures
span open.
were higher in were
Temperatures the layers
higher immediately
in the layers below the bridge
immediately due to
below thewarmer
bridge water
due tobeing
warmer forcedwaterto lower
being
layers by
forced the presence
to lower layers byof the
the bridge
presence relative
of thetobridge
the condition
relative to withthedraw span with
condition open.draw span open.

Figure 20. (a,b) Temperature difference of baseline scenario with HCB relative to test 1 without HCB
during peak ebb and flood; (c,d) velocity difference of baseline scenario with HCB relative to test 2
with draw span open during ebb andand flood.
flood. (Typical
(Typical spring
spring tide
tide on
on 27
27 April
April 2017).
2017).

4. Discussion
4. Discussion

Zone of Influence—Quantitative Assessment

Zone of Influence—Quantitative Assessment
The ZOI extends beyond the Hood Canal width scale of 2.4 km based on the results presented
The ZOI extends beyond the Hood Canal width scale of 2.4 km based on the results presented
in Section 3. However, to quantify spatial extent, environmentally significant influence must first
in Section 3. However, to quantify spatial extent, environmentally significant influence must first be
be defined. Maximum difference induced by HCB relative to test 1 on currents without the bridge
defined. Maximum difference induced by HCB relative to test 1 on currents without the bridge
occurred at the bridge itself and decayed or dissipated with distance from the bridge. For the purpose
occurred at the bridge itself and decayed or dissipated with distance from the bridge. For the purpose
of this assessment, we defined ZOI as the distance at which the maximum difference (∆) dropped to
of this assessment, we defined ZOI as the distance at which the maximum difference (Δ) dropped to
10% of its value for current and salinity parameters (e.g., if maximum ∆ = 1 m/s, then ZOI is where
10% of its value for current and salinity parameters (e.g., if maximum Δ = 1 m/s, then ZOI is where Δ
∆ = 0.1 m/s).
= 0.1 m/s).
We first assessed the horizontal extent of ZOI by examining the changes in surface currents
We first assessed the horizontal extent of ZOI by examining the changes in surface currents
induced by the bridge relative to test 1. With a focus on impacts during peak tidal currents, we averaged
induced by the bridge relative to test 1. With a focus on impacts during peak tidal currents, we
multiple (71) peak ebb and peak flood instances over the calibration simulation period (25 April–11 June
averaged multiple (71) peak ebb and peak flood instances over the calibration simulation period (25
2017). We then examined the vertical longitudinal extent of the ZOI by examining the relative difference
April–11 June 2017). We then examined the vertical longitudinal extent of the ZOI by examining the
of the variables for all depth layers along Transect-b over a distance of 20 km. Figure 21 shows the
relative difference of the variables for all depth layers along Transect-b over a distance of 20 km.
average difference in peak currents due to the presence of HCB relative to test 1 for the 10 model
Figure 21 shows the average difference in peak currents due to the presence of HCB relative to test 1
layers. The distance between −10 km and 0 km represents the region south of the bridge, while that
for the 10 model layers. The distance between −10 km and 0 km represents the region south of the
between 0 km and 10 km represents the region north of the bridge. Model layers were distributed
bridge, while that between 0 km and 10 km represents the region north of the bridge. Model layers
were distributed using a power law with an exponent of 1.5 such that layers had a higher
J. Mar. Sci. Eng. 2018, 6, 119 18 of 22

using a power
J. Mar. Sci. Eng. law
2018, with
6, x an exponent of 1.5 such that layers had a higher concentration near the18surface.
of 22
Surface layer 1 occupied 3% of the water column and the bottom layer 10 occupied 15% of the water
concentration
column. near thebased
For currents, surface.
onSurface layermultiple
averaging 1 occupied 3% of
peak theand
ebb water column
peak floodand the bottom
instances, layer
maximum
difference (−0.70 m/s) occurred during the ebb and in the surface layer, extending ~2.02 kmpeak
10 occupied 15% of the water column. For currents, based on averaging multiple peak ebb and north
flood instances, maximum difference (−0.70 m/s) occurred during the ebb and in the surface layer,
before dropping to −0.07 m/s (10% of the maximum difference). The maximum difference during
extending ~2.02 km north before dropping to −0.07 m/s (10% of the maximum difference). The
the flood (−0.57 m/s) also occurred in the surface layer and the influence extended up to 3.43 km.
maximum difference during the flood (−0.57 m/s) also occurred in the surface layer and the influence
It is interesting to note that, while there was a reduction in surface currents, the model generated
extended up to 3.43 km. It is interesting to note that, while there was a reduction in surface currents,
a compensating increase in bottom-layer currents during flood, as well as ebb. The zone of influence,
the model generated a compensating increase in bottom-layer currents during flood, as well as ebb.
therefore, extends over the entire water column. It is important to note that this ZOI for currents is
The zone of influence, therefore, extends over the entire water column. It is important to note that
based
this on
ZOIaverage impacts
for currents during
is based onpeak currents.
average impactsThe ZOI spatial
during extent is
peak currents. smaller
The at all other
ZOI spatial extenttimes
is
with smaller-magnitude currents.
smaller at all other times with smaller-magnitude currents.

Figure
Figure 21.21.Difference
Differencein
in current
current magnitudes
magnitudesdue duetoto
HCB
HCBrelative to test
relative 1, plotted
to test alongalong
1, plotted Transect-b for
Transect-b
forall
allmodel
model layers. TheThe
layers. (top) panel
(top) presents
panel peak peak
presents ebb and the
ebb (bottom)
and panel presents
the (bottom) peak flood
panel presents results
peak flood
results averaged over the calibration period of 25 April–11 June 2017. (The distance between −10 0km
averaged over the calibration period of 25 April–11 June 2017. (The distance between −10 km and
andkm0 represents
km representsthe region southsouth
the region of theof
bridge, while that
the bridge, between
while 0 km and
that between 10 km
0 km represents
and the regionthe
10 km represents
north of the bridge).
region north of the bridge).

Furtherexamination
Further examinationofofthe theresults
results showed
showed that,
that, unlike
unlikevelocity,
velocity,peak
peakinstantaneous
instantaneous difference
difference
between baseline with bridge and test 1 without the bridge for temperature
between baseline with bridge and test 1 without the bridge for temperature and salinity did not and salinity didalways
not
always
occur occur
at the bridgeat the
and bridge
occurredand at occurred at various
various times in the times in thePeak
tidal cycle. tidalinstantaneous
cycle. Peak instantaneous
differences that
vary in space and time make it difficult to define the ZOI. The approach forZOI.
differences that vary in space and time make it difficult to define the The approach
temperature for
and salinity
temperature and salinity was, therefore, to develop an average ZOI,
was, therefore, to develop an average ZOI, taking into consideration all time steps. taking into consideration all time
steps.
Figure 22 shows the difference in salinity due to the presence of HCB relative to test 1 computed
Figure 22 shows the difference in salinity due to the presence of HCB relative to test 1 computed
by averaging the differences over all hourly time steps for each layer. The average salinity difference
by averaging the differences over all hourly time steps for each layer. The average salinity difference
results show a consistent pattern north and south of the bridge. The peak influence in average salinity
results show a consistent pattern north and south of the bridge. The peak influence in average salinity
occurred at the bridge, and, as described in Section 3, it was controlled by the mixing induced by the
occurred at the bridge, and, as described in Section 3, it was controlled by the mixing induced by the
structure
structureduring
duringebbebband
andflood.
flood.This
This resulted
resulted in an increase
in an increasein insalinity
salinityininthe
theupper
upper layers
layers above
above thethe
bridge
bridgedraft depth
draft depth(0–4.57
(0–4.57m) m)and
andaacorresponding reductionin
corresponding reduction insalinity
salinityininthe
thelower
lower layers
layers below
below thethe
draft
draft from 4.57 m to ~20 m. The near-bed layers were unaffected. Also noticeable is the extent of thethe
from 4.57 m to ~20 m. The near-bed layers were unaffected. Also noticeable is the extent of
effect north
effect of the
north bridge
of the bridgein the surface
in the where
surface wherethethe
surface
surfacelayers show
layers showelevated
elevated salinity levels
salinity ~+0.1
levels ~+0.1PSU
that
PSUextend
that all the way
extend to the
all the waymouth
to theofmouth
Hood ofCanal.
HoodThe effects
Canal. Theofeffects
bridge-induced mixing resulted
of bridge-induced mixing in
higher salinity
resulted concentrations
in higher in the surfacein
salinity concentrations layers that were
the surface transported
layers that werenorth in the surface
transported north inoutflow
the
direction. This resulted in persistent and elevated levels of background salinity
surface outflow direction. This resulted in persistent and elevated levels of background salinity difference north of the
difference north of the bridge of 0.06 PSU over 4–10 km. The maximum difference (Δ) in salinity at
J. Mar. Sci. Eng. 2018, 6, 119 19 of 22

J. Mar. Sci. Eng. 2018, 6, x 19 of 22

J. Mar. Sci. Eng. 2018, 6, x 19 of 22
bridge of 0.06 PSU over 4–10 km. The maximum difference (∆) in salinity at the bridge was +0.42 PSU
the bridge was +0.42 PSU that extended 1.96 km to the south before dropping to 10% (0.04 PSU). ZOI
that
theextended 1.96+0.42
bridge was km to
PSU thethat
south before1.96
dropping to 10% (0.04 PSU). ZOI determination to theZOI
north
determination to the north wasextended
complex as km to the
background south before
salinity dropping
increased to 10% (0.04 PSU).
past the salinity 10% Δ
was complex as background
determination to the northsalinity increased
was complex past the salinity
as background salinity ∆ level. The
10%increased past ZOI to the north
the salinity 10% was,
Δ
level. The ZOI to the north was, therefore, provided as a range from 3.80 km where salinity Δ dropped
therefore,
level. Theprovided asnorth
ZOI to the a range from
was, 3.80 km
therefore, where salinity
provided ∆ from
as a range dropped to 10%
3.80 km above
where theΔbackground
salinity dropped
to 10% above the background level of 0.1 PSU to Hood Canal mouth at ~10 km.
level of 0.1 PSU to Hood Canal mouth at ~10 km.
to 10% above the background level of 0.1 PSU to Hood Canal mouth at ~10 km.

Figure
Figure 22.22. Differenceininsalinity
Difference salinitydue
duetotoHCB
HCB relative
relative to
to test
test 1,1,plotted
plottedalong
alongTransect-b
Transect-b forfor
allall
model
model
layers.
layers. The
The plotshows
plot showsaverage
averagesalinity
salinitydifference
difference for all
all hourly
hourlytime timesteps
stepsover
overthe
thecalibration
calibration period
period
of of
25 25 April–11
April–11 JuneJune 2017.(The
2017. (Thedistance
distancebetween
between −−10
10km
−10 kmand
km and000km
and kmrepresents
km representsthe
represents theregion south
region south of of
thethe
bridge, while that between 0 km and 10 km
bridge, while that between 0 km and 10 km represents represents the
represents the region
the region north
region north of
north ofthe
of thebridge).
the bridge).
bridge).

Figure
Figure 23 23 shows the difference in temperaturedue duetotothe thepresence
presenceofofthe thebridge
bridgerelative
relativetototesttest11for
Figure 23 shows
showsthe thedifference
differenceinintemperature
temperature due to the presence of the bridge relative to test 1
thefor the
10 model 10 model
layers. layers.
The The The effect
effecteffect
of HCB of HCB on temperature,
on temperature, based
basedbased on averaging
on averaging the difference
the difference for all
for all hourly
for the 10 model layers. of HCB on temperature, on averaging the difference for all
hourly
time steps time
over steps over the
the period of 25 period of 25 April–11
April–11 Junesimilar
2017, was similar to thatThefor influence
salinity. The
hourly time steps over the period of 25June 2017, was
April–11 June 2017, to that
was for salinity.
similar to that for salinity.ofThethe
influence
bridge on of of the
near-field bridge on
temperature,near-field
due temperature,
to the mixing due to the
induced mixing induced
by the bridge, by the bridge, resulted
influence the bridge on near-field temperature, due to the mixing inducedresulted in a decrease
by the bridge, resulted in
in a layer
upper decrease in upper layer
temperatures and temperatures
an increase inand an increaseininlower
temperature temperature
layers in lower
below the layers
draft below
from the m
4.57
in a decrease in upper layer temperatures and an increase in temperature in lower layers below the
draft from 4.57 m to ~20 m. The near-bed layers were unaffected. The effect of the bridge on
to ~20from
draft m. The 4.57near-bed layers were unaffected. The effect of the bridge onThe temperature is noticeable in
temperature is m to ~20
noticeable m. Thedownstream
in the near-bed layers regionwere northunaffected.
of the bridge where effecttheofsurface
the bridge
layerson
the downstream
temperature region
is noticeable north of the bridge
in the downstream where the surface
region north layers showed a reduction in temperature
showed a reduction in temperature levels that extended all theofway
thetobridge
the mouthwhere the surface
of Hood Canal. Thelayers
levels
showed that
a extendedinall
reduction the way tolevels
temperature the mouth
that of HoodallCanal.
extended the way Theto maximum
the mouth average
of Hood difference
Canal. The
maximum average difference (−0.49 °C) in the surface layers, a reduction in surface temperatures,
−0.49 ◦ C) average
(maximum in the surface layers,
difference a reduction in surface temperatures, dissipated to 10% ∆T (0.05 ◦ C)
dissipated to 10% ΔT (0.05 °C)(−0.49
~2.23 km°C)to inthe
thesouth.
surface layers, as
However, a reduction
in the case of in salinity,
surface temperature
temperatures,
~2.23 km totothe
dissipated 10% south.
ΔT to However,
(0.05 as inkm thetocase
the of salinity, temperature ZOI determination to the north
ZOI determination the°C) ~2.23
north was complex south.
as thereHowever, as in the
was a decrease incase of salinity,
background temperature
temperature
was complex as there was a decrease in background temperature difference of 0.11 ◦ C, over 4–10 km,
ZOI determination
difference of 0.11 °C,to over
the north
4–10 km,wasthatcomplex
exceeded as the
there was a decrease
temperature 10% ΔTin background
level. The ZOI totemperature
the north
that exceeded
difference
was, therefore, the temperature
of 0.11 provided
°C, over 4–10 10% ∆T
km, that
as a range level.
from The
exceeded
4.51 kmthe ZOI
where to the north
temperature
temperature was,
10%ΔT therefore,
ΔTdropped provided
level. The ZOI°C,
to 0.16 toas
the
10%a range
north
Δ
from 4.51 km where temperature ∆T dropped to 0.16 ◦ C, 10% ∆ above the background level, to Hood
above
was, the background
therefore, providedlevel, to Hood
as a range from Canal
4.51mouth
km whereat ~10temperature
km. It is noted, however, to
ΔT dropped that
0.16 the°C,
bridge
10% Δ
Canal
above mouth
causedthean at ~10 km.
increase
background It is noted,
in temperatures
level, to Hood however,
in Canal that
the lowermouth the at
layersbridge
of
~10 caused
upkm.to 0.20 an
It is°C. increase
noted, in temperatures
however, that the bridgein the
lower layers of up to 0.20 ◦ C.
caused an increase in temperatures in the lower layers of up to 0.20 °C.

Figure 23. Difference in temperature due to HCB relative to test 1, plotted along Transect-b for all
model layers. The plot shows average temperature difference for all hourly time steps over the
Figure 23. Difference
23. Difference
calibration period ofinin
25 temperature
temperature due
due
April–11 June to to
HCB
2017. HCB relative
relative
(The to 1,
to test
distance test
between 1,−10
plotted
plotted along
km andalong
0 kmTransect-b
Transect-b for
the all
for all model
represents
layers.
model The
region plotThe
layers.
south ofshows
plot
the average
shows
bridge, temperature
average
while difference
temperature
that between for10
allkm
difference
0 km and hourly
for time
all
represents steps
hourly over steps
time
the region the calibration
north over
of thethe
period of 25period
calibration
bridge). April–11 June
of 25 2017. (The
April–11 Junedistance between
2017. (The −10between
distance km and −100 kmkm represents
and 0 km the region south
represents the
of the bridge,
region south ofwhile
the that
bridge,between
while0that
km and 10 km0 represents
between km and 10the kmregion north the
represents of the bridge).
region north of the
bridge).
J. Mar. Sci. Eng. 2018, 6, 119 20 of 22

Table 5 summarizes the ZOI dimensions for the parameters of interest examined using the criterion
of a reduction in maximum difference ∆ to 10% of its value.

Table 5. Summary of zone of influence (ZOI) characterizing distances over which the difference in
predicted maximum difference ∆ due to the presence of the bridge relative to the test 1 scenario without
the bridge was reduced to 10% of its value. Results in this table are based on an average of all peak ebb
and flood instances over the period of 25 April–11 June 2017 for velocity, and an average of all time
steps for salinity and temperature.

South HCB North HCB

Variable Max. (∆)
ZOI (km) ZOI (km)
Velocity (m/s) −0.70 2.02 3.43
Salinity (PSU) 0.42 1.96 3.80–10
Temperature (◦ C) −0.49 2.23 4.51–10

Although the ZOI calculations and the values presented in Table 5 above used average values of
maximum differences, it is noted that instantaneous differences could be as high as 1.22 m/s, 1.66 PSU,
and 2.16 ◦ C for velocity, salinity, and temperature, respectively.

5. Conclusions
A combination of oceanographic field data measurements and modeling-based assessment
were used to ascertain whether the presence of HCB impacts near-field oceanographic parameters.
While some effect was expected simply due to the existence of a large floating object in the path of
the tidally influenced fjord-like waterbody of Hood Canal, the extent of influence was not quantified
in the past. The results were conclusive in that both field data and model results showed that HCB
blocks the movement of surface layer currents over ~85% of the width of Hood Canal. The openings at
each end of the bridge (15% of the Hood Canal width) do carry a portion of the tidal flow; however,
these openings are in shallow depths and do not fully compensate for the surface layer blockage.
As a result, the current profile structure in the water column is altered. The reduction in surface layer
currents and tidal flow is compensated by a corresponding increase in bottom-layer currents and flow.
The impact is also felt in salinity and temperature magnitudes, particularly in the upper 20 m of the
water column.
The ZOI was defined as the distances over which the difference in predicted maximum difference
due to the presence of the bridge relative to the conditions without the bridge was reduced to 10%
of its value. Based on the simulation conducted using 2017 spring conditions, the ZOI for current
extends 3.43 km to the north and 2.02 km to the south of HCB. Maximum current difference based
on the average of peak currents was −0.70 m/s at the surface layer. Similarly, the ZOI for salinity
extends 3.80 km to the north and 1.96 km to the south, over which salinity deviated from the condition
without the bridge. Maximum salinity difference was an increase of 0.42 PSU in the surface layers and
an increase of 0.26 PSU in layers immediately below the bridge, based on a temporal average over the
duration of the simulation. The ZOI for temperature extends 2.23 km to the south of HCB, over which
temperature deviated from the condition without the bridge. Maximum temperature difference at the
bridge was a decrease 0.49 ◦ C in the surface layers and a decrease in temperature of 0.36 ◦ C in layers
immediately below the bridge, based on a temporal average over the duration of the simulation.
The near-field influence of the bridge on physical parameters of currents, salinity, and temperature
as quantified by ZOI extends over 1.96 km to 5.15 km, which is roughly one to two Hood Canal width
scales. Our study concludes that the influence on temperature and salinity is restricted to the upper
20 m, but extends through the water depth for current. The influence on temperature and salinity
persists beyond the ZOI all the way to the mouth of Hood Canal at the confluence with Admiralty
Inlet. The downstream effect is associated with fjord-like estuarine circulation. The brackish outflow in
Hood Canal is restricted to surface layers. The presence of the bridge affects this net outflow through
J. Mar. Sci. Eng. 2018, 6, 119 21 of 22

blockage and added turbulence and mixing. The persistent impact is seen in salinity and temperature
downstream of the bridge, even after tides are averaged out.

Author Contributions: T.K. led the investigation and was responsible for conceptualization, methodology, and
model development and application. T.W. developed the bridge module software. A.N. was responsible for
processing and analysis of model results and field data. T.K. and A.N. analyzed the simulation results and
developed the methodology for quantifying and summarizing the results.
Funding: This research was managed by Long Live the Kings through grants from the State of Washington.
Funding was provided by Washington State with equal in-kind contributions by those participating in the research.
Acknowledgments: This publication is part of the Hood Canal Bridge Ecosystem Impact Assessment:
an ecosystem-wide assessment of potential impacts of the Hood Canal Bridge on fish passage and water quality
in Hood Canal.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to
publish the results.

References
1. Puget Sound Partnership. 2009 State of the Sound Report; No. PSP09-08; Puget Sound Partnership: Olympia,
WA, USA, 2010.
2. Khangaonkar, T.; Yang, Z.; Kim, T.; Roberts, M. Tidally Averaged Circulation in Puget Sound Sub-basins:
Comparison of Historical Data, Analytical Model, and Numerical Model. J. Estuar. Coast. Shelf Sci. 2011, 93,
305–319. [CrossRef]
3. Khangaonkar, T.; Sackmann, B.; Long, W.; Mohamedali, T.; Roberts, M. Simulation of annual biogeochemical
cycles of nutrient balance, phytoplankton bloom(s), and DO in Puget Sound using an unstructured grid
model. Ocean Dyn. 2012, 62, 1353–1379. [CrossRef]
4. Khangaonkar, T.; Nugraha, A.; Xu, W.; Long, W.; Bianucci, L.; Ahmed, A.; Mohamedali, T.; Pelletier, G.
Analysis of Hypoxia and Sensitivity to Nutrient Pollution in Salish Sea. J. Geophys. Res. Oceans 2018, 123,
4735–4761. [CrossRef]
5. Newton, J.; Bassin, C.; Devol, A.; Richey, J.; Kawase, M.; Warner, M. Chapter 1: Overview and Results
Synthesis. In Integrated Assessment and Modeling Report; Hood Canal Dissolved Oxygen Program; University
of Washington: Washington, DC, USA, 2011. Available online: http://www.hoodcanal.washington.edu/
documents/document.jsp?id=3019 (accessed on 10 December 2018).
6. Cope, B.; Roberts, M. Review and Synthesis of Available Information to Estimate Human Impacts to Dissolved
Oxygen in Hood Canal; Ecology Publication No. 13-03-016; EPA Publication No. 910-R-13-002; Washington
State Department of Ecology and U.S. EPA: Washington, DC, USA, 2013. Available online: https://fortress.
wa.gov/ecy/publications/documents/1303016.pdf (accessed on 10 December 2018).
7. Khangaonkar, T.; Wang, T. Potential alteration of fjordal circulation due to a large floating structure—
Numerical investigation with application to Hood Canal basin in Puget Sound. Appl. Ocean Res. 2013, 39,
146–157. [CrossRef]
8. Moore, M.; Berejikian, B.A.; Tezak, E.P. A floating bridge disrupts seaward migration and increases mortality
of steelhead smolts in Hood Canal, Washington State. PLoS ONE 2013, 8, e73427. [CrossRef] [PubMed]
9. Hood Canal Bridge Assessment Team. Hood Canal Bridge Ecosystem Impact Assessment Plan: Framework and
Phase 1 Details; Long Live the Kings: Seattle, WA, USA, 2016. Available online: https://lltk.org/wp-content/
uploads/2016/08/Hood-Canal-Bridge-Impact-Assessment-Plan-FINAL-27September2016.pdf (accessed on
10 December 2018).
10. RPS. Current Measurements in Hood Canal Data Report; Maya Whitmont and Kevin Redmond of RPS Group
PLC: Seattle, WA, USA, 2017.
11. Chen, C.; Liu, H.; Beardsley, R.C. An unstructured, finite-volume, three-dimensional, primitive equation
ocean model: Application Tocoastal Ocean and estuaries. J. Atmos. Ocean. Technol. 2003, 20, 159–186.
[CrossRef]
12. Cerco, C.; Cole, T. Three-Dimensional Eutrophication Model of Chesapeake Bay; Tech. Rep. EL-94-4; U.S. Army
Corps of Engineers: Vicksburg, MI, USA, 1994.
J. Mar. Sci. Eng. 2018, 6, 119 22 of 22

13. Cerco, C.F.; Cole, T.M. User’s Guide to the CE-QUAL-ICM Three-Dimensional Eutrophication Model; Release
Version 1.0; U.S. Army Corps of Engineers: Washington, DC, USA, 1995; p. 320.
14. Khangaonkar, T.; Long, W.; Xu, W. Assessment of circulation and inter-basin transport in the Salish Sea
including Johnstone Strait and Discovery Islands pathways. Ocean Model. 2017, 109, 11–32. [CrossRef]
15. Bianucci, L.; Long, W.; Khangaonkar, T.; Pelletier, G.; Ahmed, A.; Mohamedali, T.; Roberts, M.;
Figueroa-Kaminsky, C. Sensitivity of the regional ocean acidification and the carbonate system in Puget
Sound to ocean and freshwater inputs. Elem. Sci. Anthropocene 2018, 6, 22. [CrossRef]
16. Mohamedali, T.; Roberts, M.; Sackmann, B.S.; Kolosseus, A. Puget Sound Dissolved Oxygen Model: Nutrient
Load Summary for 1999–2008; Publication No. 11-03-057; Washington State Department of Ecology: Olympia,
WA, USA, 2011.
17. Spargo, E.; Westerink, J.; Luettich, R.; Mark, D. Developing a Tidal Constituent Database for the Eastern
North Pacific Ocean. In Proceedings of the 8th International Conference on Coastal Modeling, Mexico City,
Mexico, 18–24 February 2007.
18. Locarnini, R.A.; Mishonov, A.V.; Antonov, J.I.; Boyer, T.P.; Garcia, H.E.; Baranova, O.K.; Zweng, M.M.;
Paver, C.R.; Reagan, J.R.; Johnson, D.R.; et al. World Ocean Atlas 2013; Volume 1: Temperature; NOAA Atlas
NESDIS 73; Levitus, S., Mishonov, A., Eds.; National Oceanic and Atmospheric Administration: Silver
Spring, MD, USA, 2013; p. 40.
19. Zweng, M.M.; Reagan, J.R.; Antonov, J.I.; Locarnini, R.A.; Mishonov, A.V.; Boyer, T.P.; Garcia, H.E.;
Baranova, O.K.; Johnson, D.R.; Seidov, D.; et al. World Ocean Atlas 2013; Volume 2: Salinity; NOAA Atlas
NESDIS 74; Lev-itus, S., Mishonov, A., Eds.; National Oceanic and Atmospheric Administration: Silver
Spring, MD, USA, 2013; p. 39.
20. Willmott, C.J. Some comments on the evaluation of model performance. Bull. Am. Meteorol. Soc. 1982, 63,
1309–1313. [CrossRef]
21. Wang, T.; Khangaonkar, T.; Long, W.; Gill, G. Development of a Kelp-Type Structure Module in a Coastal
Ocean Model to Assess the Hydrodynamic Impact of Seawater Uranium Extraction Technology. J. Mar. Sci.
Eng. 2014, 2, 81–92. [CrossRef]
22. Wissmar, R.; Simenstad, C. Energetic Constraints of Juvenile Chum Salmon (Oncorhynchus keta) Migrating in
Estuaries. Can. J. Fish. Aquat. Sci. 1988, 45, 1555–1560. [CrossRef]
23. Greene, C.M.; Hall, J.; Small, D.; Smith, P. Effects of Intertidal Water Crossing Structures on
Estuarine Fish and Their Habitat: A Literature Review and Synthesis. 2017. Available online:
http://www.skagitclimatescience.org/wp-content/uploads/2018/07/Greene-et-al.-2017-review-
on-intertidal-water-crossing-structures-and-fish-1.pdf (accessed on 10 December 2018).

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).